tdavidson/hate_speech_offensive
收藏资源简介:
该数据集是一个用于检测仇恨言论和冒犯性语言的标注数据集,主要基于推文数据。数据集的语言为英语,数据规模在10K到100K之间,属于单语种数据集。数据集的结构包括多个字段,如推文内容、仇恨言论标注、冒犯性语言标注等。数据集的创建过程涉及专家生成和众包标注,但具体的标注过程和来源数据的信息未详细说明。数据集的使用可能涉及社会影响和偏见问题,但具体讨论未提供。
This is an annotated dataset developed for detecting hate speech and offensive language, primarily sourced from tweet data. It is a monolingual English dataset with a scale ranging from 10K to 100K samples. The dataset structure includes multiple fields such as tweet content, hate speech annotation labels, and offensive language annotation labels. The dataset's creation process incorporates both expert generation and crowdsourced annotation, yet specific details regarding the annotation procedure and source data remain unspecified. Potential social impact and bias-related issues may emerge during the usage of this dataset, but no specific discussion on these aspects has been offered.
数据集概述
- 名称: Hate Speech and Offensive Language
- 语言: 英语 (
en) - 许可: 未知
- 多语言性: 单语种
- 大小: 10K<n<100K
- 来源: 原始数据
- 任务类别: 文本分类
- 标签: 仇恨言论检测
数据集结构
数据实例
json { "count": 3, "hate_speech_annotation": 0, "offensive_language_annotation": 0, "neither_annotation": 3, "label": 2, # "neither" "tweet": "!!! RT @mayasolovely: As a woman you shouldnt complain about cleaning up your house. & as a man you should always take the trash out..." }
数据字段
- count: 整数,标记每个推文的用户数(最小为3)
- hate_speech_annotation: 整数,标记为仇恨言论的推文用户数
- offensive_language_annotation: 整数,标记为攻击性语言的推文用户数
- neither_annotation: 整数,标记为既非攻击性也非非攻击性的推文用户数
- label: 类别标签,多数用户的大类标签(0: hate-speech, 1: offensive-language 或 2: neither)
- tweet: 字符串,推文内容
数据分割
- 训练集: 24783个样本,数据大小为3207814字节
数据集创建
注释
- 创建者: 专家生成和众包
- 个人和敏感信息: 用户名未匿名化
使用数据集的考虑
许可证信息
- 许可证: MIT License
引用信息
bibtex @inproceedings{hateoffensive, title = {Automated Hate Speech Detection and the Problem of Offensive Language}, author = {Davidson, Thomas and Warmsley, Dana and Macy, Michael and Weber, Ingmar}, booktitle = {Proceedings of the 11th International AAAI Conference on Web and Social Media}, series = {ICWSM 17}, year = {2017}, location = {Montreal, Canada}, pages = {512-515} }




